R help - Nov 2008 - Unexpected behavior of clocktime related to daylight savings time

Colleagues,

I submitted this several days ago and no one responded, so I am trying  
again, trying a different subject line:

I just encountered some unexpected behavior of difftime in  
relationship to the change from daylight savings to standard time.

My understanding is that DST and ST take effect at 2AM.  However, the  
results below suggests that R (version 2.8.0 in OS X) implements the  
change at 2:16AM:

Expected:
> > difftime("2008-11-02 02:01:00", "2008-11-02
00:59:00")
> Time difference of 2.033333 hours
> > difftime("2008-11-02 01:16:00", "2008-11-02
01:15:00")
> Time difference of 1 mins
> > difftime("2008-11-02 01:18:00", "2008-11-02
01:17:00")
> Time difference of 1 mins
Not expected:
> > difftime("2008-11-02 01:17:00", "2008-11-02
01:16:00")
> Time difference of 1.016667 hours
Can anyone explain this?

Dennis


Dennis Fisher MD
P < (The "P Less Than" Company)
Phone: 1-866-PLessThan (1-866-753-7784)
Fax: 1-415-564-2220
www.PLessThan.com


	[[alternative HTML version deleted]]

On Fri, 7 Nov 2008, Dennis Fisher wrote:
> Colleagues,
>
> I submitted this several days ago and no one responded, so I am trying
> again, trying a different subject line:
Well, you posted something that indicated you had not studied the relevant 
help pages, without the information requested in the R posting guide, and 
with an HTML posting.
> I just encountered some unexpected behavior of difftime in
> relationship to the change from daylight savings to standard time.
>
> My understanding is that DST and ST take effect at 2AM.  However, the
> results below suggests that R (version 2.8.0 in OS X) implements the
> change at 2:16AM:
The transition time depends on the country (and in some cases, the year).
In the EU it is at 2am (and always has been, not that the EU is very old).
> Expected:
>>> difftime("2008-11-02 02:01:00", "2008-11-02
00:59:00")
>> Time difference of 2.033333 hours
>>> difftime("2008-11-02 01:16:00", "2008-11-02
01:15:00")
>> Time difference of 1 mins
>>> difftime("2008-11-02 01:18:00", "2008-11-02
01:17:00")
>> Time difference of 1 mins
>
> Not expected:
>>> difftime("2008-11-02 01:17:00", "2008-11-02
01:16:00")
>> Time difference of 1.016667 hours
>
> Can anyone explain this?
Yes, and you should have been able to do so from the information in the 
help pages.  As ?strptime says

      Remember that in most timezones some times do not occur and some
      occur twice because of transitions to/from summer time.  What
      happens in those cases is OS-specific.

See also the note on ?Sys.timezone.  No OS I tried did this, not even Mac 
OS X set to EST5EDT.  But in EST5EDT 2008-11-02 01:17:00 occurred twice:
> as.POSIXct("2008-11-02 01:17:00")
[1] "2008-11-02 01:17:00 EDT"
> as.POSIXct("2008-11-02 01:17:00") + 3600
[1] "2008-11-02 01:17:00 EST"

so 1.016667 hours would be one of two correct answers in that timezone.

However, you didn't tell us what timezone you are in and hence we cannot 
know when DST transitions occur in that timezone.
>
> Dennis
>
>
> Dennis Fisher MD
> P < (The "P Less Than" Company)
> Phone: 1-866-PLessThan (1-866-753-7784)
> Fax: 1-415-564-2220
> www.PLessThan.com
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

I'm not surprised that strange things happen if you try to handle times in 
the duplicated hour without specifying which time you mean.  If your 
current time zone setting is something like EST5EDT that includes both 
daylight-saving and standard times then  2008-11-02 01:16:00 occured 
twice, so some of the calls to difftime() are ambiguous.  If your 
time-zone setting specifies daylight-saving time or not, then some of the 
calls to difftime() involve times that did not exist.  If you create 
POSIXt objects that unambiguously specify the times then things should be 
more predictable.

 	-thomas

On Fri, 7 Nov 2008, Dennis Fisher wrote:
> Colleagues,
>
> I submitted this several days ago and no one responded, so I am trying
> again, trying a different subject line:
>
> I just encountered some unexpected behavior of difftime in
> relationship to the change from daylight savings to standard time.
>
> My understanding is that DST and ST take effect at 2AM.  However, the
> results below suggests that R (version 2.8.0 in OS X) implements the
> change at 2:16AM:
>
> Expected:
>>> difftime("2008-11-02 02:01:00", "2008-11-02
00:59:00")
>> Time difference of 2.033333 hours
>>> difftime("2008-11-02 01:16:00", "2008-11-02
01:15:00")
>> Time difference of 1 mins
>>> difftime("2008-11-02 01:18:00", "2008-11-02
01:17:00")
>> Time difference of 1 mins
>
> Not expected:
>>> difftime("2008-11-02 01:17:00", "2008-11-02
01:16:00")
>> Time difference of 1.016667 hours
>
> Can anyone explain this?
>
> Dennis
>
>
> Dennis Fisher MD
> P < (The "P Less Than" Company)
> Phone: 1-866-PLessThan (1-866-753-7784)
> Fax: 1-415-564-2220
> www.PLessThan.com
>
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle